Supplemental Information

Supplemental Methods CDNA Constructs and Vectors

Full length FLAG tagged murine HoxA9 (1) (99% homology to human) was PCR

amplified from PRC-CMV-FLAG HoxA9 (gift, C.Largman, UCSF, San Francisco, CA.)

using forward T7 and reverse mur HoxA9-Not1 5' GAT CGC GGC CGC TAA GCC

CAA ATG GCA TCA 3' primers and subsequently subcloned into the SalI-NotI vector

fragment of the Hermes HRS puro IRES eGFP retroviral plasmid (2) (gift, H.Blau,

Stanford University, Stanford, CA). The SalI-NotI FLAG tagged HoxA9 fragment from

Hermes HRS puro Hox9 IRES eGFP was replaced with an HA tagged murine HoxA9

PCR product amplified from pHRS-puro-Flag-HoxA9-ires-eGFP using forward primer

BamH1-Sal1-HA-mur-HoxA9 5' GCG GGA TCC GTC GAC CCA CCA TGG GCT

ACC CCT ACG ACG TGC CCG ACT ACG CCA TGG CCA CCA CCG GGG CCC T

3' and reverse primer mur HoxA9-Not1. pcDNA3.1 HA HoxA9 was derived by cloning

the HA tagged murine HoxA9 PCR product into the BamHI-NotI vector fragment of

pcDNA3.1 (Invitrogen). Full length wild-type HA tagged BRCA1 (3) (gift, F.Rauscher,

Wistar Institute, Philadelphia, PA) in the pcDNA3.1 vector was partially digested with

BamHI-KpnI or KpnI-NotI to obtain the HA tagged 5’ end or 3’ end of BRCA1. Hermes

HRS puro IRES eGFP was partially digested with NotI-XbaI to obtain the IRES-eGFP

fragment. All three fragments were ligated into the Hermes HRS puro IRES eGFP

BamHI-NotI vector fragment. pcDNA3.1 HA tagged BRCA1 Δ exon 11b (3) (gift,

F.Rauscher, Wistar Institute, Philadelphia, PA) was digested with BamHI-NotI and

cloned into the BamHI-XbaI vector fragment of Hermes HRS neo IRES eGFP together

S1

with the NotI-XbaI IRES-eGFP fragment (described above). The pGL2 BRCA1

luciferase plasmid (4) (gift, L.A. Chodosh, UPENN, Philadelphia, PA) was used directly.

The pGL2 BRCA1 luciferase mutants were generated by PCR amplification using Pfu

turbo polymerase (Stratagene) and the following primer pairs: Δ-223 to +44 forward 5’

GCG CGA TAT CTG CCT GCC CTC TAG CCT CTA CTC TTC 3’ and Δ-223 to +44

reverse 5’GCG CGA TAT CCG GGG GAC AGG CTG TGG GGT TTC TCA 3’, Δ-221

to -218 forward 5’ GCG CGA TAT CGC AAA CTC AGG TAG AAT TCT TCC TC 3’

and Δ-221 to -218 reverse 5’ GCG CGA TAT CCT GCC CTC TAG CCT CTA CTC

TTC CAG 3’, Δ-175 to -172 forward 5’ GCG CGA TAT CTC ATC CGG GGG CAG

ACT GGG TGG CCA 3’ and Δ-175 to -172 reverse 5’GCG CGA TAT CAA GAG ACG

GAA GAG GAA GAA TTC TAC 3’, Δ –12 to -9 forward 5’ GCG CGA TAT CGA TAA

ATT AAA ACT GCG ACT GCG CGG 3’ and Δ –12 to -9 reverse 5’GCG CGA TAT

CGC GCT TTT CCG TTG CCA CGG AAA CCA 3’. pcDNA3.1-SEAP was generated

by cloning the EcoRI-XbaI SEAP fragment from pGRE-SEAP (Clontech) into the EcoRI-

XbaI vector fragment of pcDNA3.1-eGFP (Invitrogen). CMV-PBX1 was used directly

(5). pcDNA3.1 HA-HoxA9 DNA binding mutant was generated by PCR amplification

using Pfu turbo polymerase (Stratagene) and the following primer pair: 5’GGC AGG

TCA AGA TCT GGT TCC AGA CCC GCA GGA TGA AAA TGA AGA AAA TCA 3’

and 5’ATT TTC TTC ATT TTC ATC CTG CGG GTC TGG AAC CAG ATC TTG ACC

TGC CTT TC 3’. HoxA10 cDNA (gift, H.J.Lawrence, UCSF, San Francisco, CA) was

excised from pBluescript and subcloned into the EcoR1 restriction site of the pLXSN

(Clontech) retroviral vector. Orientation was confirmed by Big DyeTM terminator

analysis (PE Biosystems) at the UCSF Biomolecular Core facility. pLKO.1-puro-

S2

luciferase shRNA and BRCA1 shRNA lentiviral plasmids were used directly (Sigma-

Aldrich, MISSIONTM TRC-Hs1.0)(6). The following MISSIONTM human BRCA1

shRNA clones were screened: TRC0000039833 (#1), TRC0000039834 (#2),

TRC0000039835 (#3), TRC0000039836 (#4), TRC0000039837 (#5). The following

MISSIONTM murine HoxA9 shRNA clones were screened: TRCN0000012508 (#1),

TRCN0000012509 (#2), TRCN0000012510 (#3), TRCN0000012511 (#4),

TRCN0000012512 (#5). The pMD2.G and pCMVΔR8.91 packaging plasmids were used

directly (gift, D.Trono, EPFL, Lausanne, Switzerland). All plasmids were confirmed by

restriction and sequence analysis (7).

Lentiviral Infection

Lentiviral particles were produced, harvested, and used to infect target cells as previously

described (8).

Expression Profiling

All experiments were performed in accordance with Institutional Review Board approval

at the University of Pennsylvania. Dissected tissues from human breast tumor and

adjacent " normal" tissue were rapidly homogenized using the Tissue TearerTM apparatus

(BioSpec Products, Inc.) and log phase cultured breast cells were harvested and total

RNA from samples was isolated and labeled, and cRNA was prepared, fragmented and

hybridized to U95A arrays, essentially as recommended by the manufacturer

(GeneChipTM protocol, Affymetrix, Inc.). The microarrays were scanned and images

were assessed for quality and normalization using GeneChip Analysis Suite 5.0

S3

(Affymetrix, Inc.). Data from each microarray analysis was exported as a .DAT file into

Rosetta ResolverTM 3.0 (Rosetta Inpharmatics, Inc.) and statistically analyzed using 2D

agglomerative clustering. Using this approach, expression data were clustered for

similarity across experiments and experiments were clustered for similarity across genes.

Probe set clusters detecting transcript level differences between normal and malignant

tissue with p<0.01 as calculated by Rosetta Resolver TM in at least four of five

tumor/normal pairs were included in the list of genes that was significantly up- or down-

regulated, using normal adjacent as the background sample. Thus up-regulated genes

correspond to transcripts that are more highly expressed in tumor compared to normal

tissue, and vice versa.

Multispectral image analysis.

Immunohistochemistry slides were examined using a Leica DMRA2 microscope (Leica

Microsystems Inc.) equipped with plan apochromatic lenses. Fields containing tumor or

normal tissues were imaged at 40X magnification through a liquid crystal filter using the

Nuance Multispectral Imaging System (Cambridge Research and Instrumentation Inc.).

The spectro-microscopic system is linked to a CCD camera and a PC. The MSI system

was used at full chip resolution, without data binning. Spectral data was acquired from

420-720 nm in 20 nm increments. Spectral unmixing was accomplished by using Nuance

software v1.42 using pure spectral libraries of individual chromogens (slides stained with

only DAB, Fast red, or hematoxylin). Images were then evaluated for the presence

of BRCA1, HoxA9 or both in normal epithelium or tumor cells using unmixed images

from the Nuance system.

S4

2D Growth Curves

50,000 cells in log phase growth were plated into fifteen 35 mm polystyrene tissue

culture plastic dishes. Three dishes were trypsinized and counted every 24 hours for five

consecutive days after they were initially plated. Data was plotted as time (days) versus

average cell number and the growth rate was determined by calculating the slope of the

line during exponential growth.

Bioinformatics Analysis

The mRNA expression levels for HoxA9 were analyzed from several independent cancer

studies using Oncomine™ (www.oncomine.org) (9). Details of standard normalization

methods and statistical calculations are provided on the Oncomine™ website.

Gene expression and clinical outcome information were obtained from two

independent publicly available data sets (10-12). Clinical outcomes from the Pawitan

study (12) was obtained from data published in the Ivshina study (11). In all cases, data

for HoxA9 was culled from normalized expression data for each breast tumor sample,

and patients were divided into quartiles based on HoxA9 expression. Each data set was

analyzed separately. For the data from the van de Vijver study, distant metastasis was

analyzed as first event only. If a patient developed a local recurrence, axillary recurrence,

contra-lateral breast cancer, or a second primary cancer (except for non-melanoma skin

cancer), she was censored at that time. Any distant metastasis after the first event was not

analyzed, based on the theoretical possibility that the secondary cancers could be a source

for distant metastases. An ipsalateral supra-clavicular recurrence was considered as first

clinical evidence for metastatic disease for this analysis. Therefore, patients with

S5

ipsalateral supra-clavicular recurrence were not censored. Patients were censored at last

follow-up. Kaplan-Meier survival curves were generated using the software WINSTAT

FOR EXCEL (R. Fitch Software), and p values were calculated by log-rank analysis.

Multivariate analyses with Cox's proportional-hazards regression were performed on the

expression levels of HoxA9 and clinicopathological variables provided in the NKI data

set with SPSS 10.0 (SPSS), with patients stratified according to their local lymph node

(LN) and estrogen receptor (ER) status, the molecular subtypes of breast cancer (13) and

further grouped into quartiles based on the relative (untransformed) expression levels of

HoxA9 (10). P-values less than 0.05 were considered significant.

S6

Supplemental References

1. Shen, W.F., Rozenfeld, S., Kwong, A., Kom ves, L.G., Lawrence, H.J., and Largman, C. 1999. HOXA9 forms triple complexes with PBX2 and MEIS1 in myeloid cells. Mol Cell Biol 19:3051-3061.

2. Rossi, F.M., Guicherit, O.M., Spicher, A., Kringstein, A.M., Fatyol, K., Blakely, B.T., and Blau, H.M. 1998. Tetracycline-regulatable factors with distinct dimerization domains allow reversible growth inhibition by p16. Nat Genet 20:389-393.

3. Wilson, C.A., Payton, M.N., Elliott, G.S., Buaas, F.W., Cajulis, E.E., Grosshans, D., Ramos, L., Reese, D.M., Slamon, D.J., and Calzone, F.J. 1997. Differential subcellular localization, expression and biological toxicity of BRCA1 and the splice variant BRCA1-delta11b. Oncogene 14:1-16.

4. Thakur, S., and Croce, C.M. 1999. Positive regulation of the BRCA1 promoter. J Biol Chem 274:8837-8843.

5. Charboneau, A., East, L., Mulholland, N., Rohde, M., and Boudreau, N. 2005. Pbx1 is required for Hox D3-mediated angiogenesis. Angiogenesis 8:289-296.

6. Moffat, J., Grueneberg, D.A., Yang, X., Kim, S.Y., Kloepfer, A.M., Hinkle, G., Piqani, B., Eisenhaure, T.M., Luo, B., Grenier, J.K., et al. 2006. A lentiviral RNAi library for human and mouse genes applied to an arrayed viral high-content screen. Cell 124:1283-1298.

7. Yu, S.F., von Ruden, T., Kantoff, P.W., Garber, C., Seiberg, M., Ruther, U., Anderson, W.F., Wagner, E.F., and Gilboa, E. 1986. Self-inactivating retroviral vectors designed for transfer of whole genes into mammalian cells. Proc Natl Acad Sci U S A 83:3194-3198.

8. Zufferey, R., Dull, T., Mandel, R.J., Bukovsky, A., Quiroz, D., Naldini, L., and Trono, D. 1998. Self-inactivating lentivirus vector for safe and efficient in vivo gene delivery. J Virol 72:9873-9880.

9. Rhodes, D.R., Yu, J., Shanker, K., Deshpande, N., Varambally, R., Ghosh, D., Barrette, T., Pandey, A., and Chinnaiyan, A.M. 2004. ONCOMINE: a cancer microarray database and integrated data-mining platform. Neoplasia 6:1-6.

10. van de Vijver, M.J., He, Y.D., van't Veer, L.J., Dai, H., Hart, A.A., Voskuil, D.W., Schreiber, G.J., Peterse, J.L., Roberts, C., Marton, M.J., et al. 2002. A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999-2009.

11. Ivshina, A.V., George, J., Senko, O., Mow, B., Putti, T.C., Smeds, J., Lindahl, T., Pawitan, Y., Hall, P., Nordgren, H., et al. 2006. Genetic reclassification of histologic grade delineates new clinical subtypes of breast cancer. Cancer Res 66:10292-10301.

12. Pawitan, Y., Bjohle, J., Amler, L., Borg, A.L., Egyhazi, S., Hall, P., Han, X., Holmberg, L., Huang, F., Klaar, S., et al. 2005. Gene expression profiling spares early breast cancer patients from adjuvant therapy: derived and validated in two population-based cohorts. Breast Cancer Res 7:R953-964.

13. Sorlie, T., Perou, C.M., Tibshirani, R., Aas, T., Geisler, S., Johnsen, H., Hastie, T., Eisen, M.B., van de Rijn, M., Jeffrey, S.S., et al. 2001. Gene expression

S7

patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 98:10869-10874.

14. Xu, C.F., Brown, M.A., Chambers, J.A., Griffiths, B., Nicolai, H., and Solomon, E. 1995. Distinct transcription start sites generate two forms of BRCA1 mRNA. Hum Mol Genet 4:2259-2264.

S8

Supplemental Table 1

*IDC-Infiltrating Ductal Carcinoma

Supplemental Table 1. Tumor characteristics of matched normal-tumor pairs analyzed by global expression profiling.

Tumor ID 1 2 3 4 5 Diagnosis* IDC IDC IDC IDC IDC Max. Diameter (cm) 2.6 4.9 1.9 1.5 2.0 Nuclear Grade HIGH HIGH HIGH HIGH HIGH Histologic Grade HIGH HIGH HIGH HIGH HIGH Estrogen Receptor Status NEG ND NEG NEG NEG

S9

Supplemental Table 2

Fold Changea Gene Description Transcripts upregulated in the tumor compared to the matched normal adjacent tissue in

at least 4 out of 5 samples 2.5 PFN2 Profilin 2 3.2 TACSTD1 Tumor-associated calcium signal transducer 1 3.2 KRT7 Keratin 7 3.5 MTHFD2 Mitochondrial methylene tetrahydrofolate dehydrogenase 3.7 CCNB2 Cyclin B2 6.1 STAT1 Signal transducer and activator of transcription 1 9.4 COL11A1 Collagen, type XI, alpha 1 11 H2AFA H2A histone family, member A

12.9 MUC1 Mucin 1 48.6 S100BPP S-100 calcium binding protein B

Transcripts downregulated in the tumor compared to the matched normal adjacent tissue in at least 4 out of 5 samples

-2.5 TGFBR2 Transforming growth factor beta receptor 2 -2.8 GAS1 Growth arrest specific 1 -3.1 HOXA4 Homeobox A4 -3.1 ID1 Inhibitor of DNA binding 1 -3.5 SEMA3C Semaphorin 3C -3.7 MEOX2 Mesenchyme homeobox 2 -4.2 PECAM1 Platelet/endothelial cell adhesion molecule 1 -4.4 HOXA9 Homeobox A9 -6.1 JAM3 Junctional adhesion molecule 3 -6.5 RAPGEF Rap1 guanine-nucleotide-exchange factor -7.6 VWF von Willebrand factor -7.7 ABC1 ATP-binding cassette 1 -8.4 DUSP1 Dual specificity phosphatase 1 -9 Col17A1 Collagen, type XVII

-10.5 CXCL12 Chemokine ligand 12 (SDF1) -11 MEOX1 Mesenchyme homeobox 1

-11.8 AQP1 Aquaporin 1 -13.9 FHL1 Four and a half LIM domains 1 -14.5 ITGA7 Integrin alpha 7 -21.4 CLDN5 Claudin 5 -23.4 FABP4 Fatty acid binding protein 4 -26.9 CNN1 Calponin 1, basic, smooth muscle -28.5 ADH1C Alcohol dehydrogenase 1C -34.7 CCR5 Chemokine receptor 5 -45.2 c-fos Fos proto-oncogene

ap-value ≤0.01 Supplemental Table 2. Select genes from a Rosetta-ResolverTM generated list of transcripts significantly altered in at least 4 out of 5 sets of matched tumor and normal adjacent tissue pairs.

S10

Supplemental Table 3 Clinical parameters related to reduced HoxA9 mRNA levels in human breast cancers

n P value Reference

47 0.0000014 Richardson, et al. Cancer Cell. 2006 Feb;9(2):121-32. breast cancer vs. normal breast

10 0.002 Turashvili et al. BMC Cancer. 2007 Mar 27;7:55.

278 0.000091 Bittner, et al. https://expo.intgen.org/expo/public/ 2005/01/15

172 0.006 Soritrou et al. J Natl Cancer Inst. 2006 Feb 15;98(4):262-72.

55 0.018 Ginestier et al. Clin Cancer Res. 2006 Aug 1;12(15):4533-44.

249 0.018 Miller et al. Proc Natl Acad Sci U S A. 2005 Sep 20;102(38):13550-5.

249 0.023 Ivshina et al. Cancer Res. 2006 Nov 1;66(21):10292-301.

high grade breast cancers

60 0.03 Ma et al. Cancer Cell. 2004 Jun;5(6):607-16.

high stage breasts cancers 244 0.00047

Bittner, et al. https://expo.intgen.org/expo/public/ 2005/01/15

tumors with complete response vs. residual disease 51 0.002 Hess et al. J Clin Oncol. 2006 Sep

10;24(26):4236-44. tumors sensitive to docetaxel 24 0.003 Chang et al. Lancet. 2003 Aug

2;362(9381):362-9. tumors with lymph node involvement (N3)

194 0.005 Bittner, et al. https://expo.intgen.org/expo/public/ 2005/01/15

large (T4) tumors 285 0.021

Bittner, et al. https://expo.intgen.org/expo/public/ 2005/01/15

tumors associated with distant metastasis 189 0.03 Desmedt et al. Clin Cancer Res. 2007 Jun 1;13(11):3207-14.

< 5 year survival 159 0.038 Pawitan et al. Breast Cancer Res. 2005;7(6):R953-64.

Supplemental Table 3. Clinical parameters related to reduced HoxA9 mRNA levels in human breast cancers.

S11

Supplemental Table 4

Signal Log Ratio

Fold Changea Gene Description

Transcripts decreased after HoxA9 induction -4.2 -17.6 NBR2 Next to BRCA1 gene 2 -4.0 -16.0 TOM1 Target of Myb1 -3.7 -13.7 CDK5R1 Regulatory subunit of cyclin-dependent kinase 5 -3.7 -13.7 DMXL1 DmX-Like 1 regulatory protein -3.5 -12.3 RENT2 Nuclear export protein

Transcripts increased after HoxA9 induction 1.0 +2.0 HoxA9 Homeobox domain protein A9 1.0 +2.0 CDK9 Cyclin-dependent protein kinase 9 1.3 +2.5 MYB MYB oncogene 1.5 +2.8 NDRG2 N-myc downstream-regulated gene2 1.9 +3.7 CSN1 Alpha S1-casein 1.9 +3.7 ACVR2 Activin 2 (TGF-beta superfamily) 2.2 +4.6 CDK8 Cyclin-dependent protein kinase 8 2.4 +5.3 MUC5B Mucin 5B 2.5 +5.7 RAP2A RAS-related protein 2A 2.6 +6.0 PCDH9 Protocadherin 9 2.7 +7.3 PRKCBP2 Protein kinase C-binding protein RACK17 2.9 +8.4 PTEN Dual specificity phospatase 3.1 +9.6 WNT10B Wingless-type MMTV integration site 10B 3.1 +9.6 BMP1 Bone morphogenetic protein 1 3.2 +10.2 COL1A2 Collagen alpha-2 type I 3.3 +10.9 NEO1 Member of NCAM cell adhesion family 3.3 +10.9 MMP1 Matrix metalloproteinase 1 4.1 +16.8 MUC1 Mucin 1 4.1 +16.8 ERBB3 Epidermal growth factor receptor 3 (HER3) 4.6 +21.1 BRCA1 BReast CAncer-related gene 1

aFold change is expressed as log2 of the signal log ratio calculated by Affymetrix Analysis Suite 5.0 p-values ≤0.001

Supplemental Table 4. Selected gene expression differences following HoxA9 induction in MDA-MB-231 cells.

S12

Supplemental Legends

Supplemental Figure S1. HoxA9 expression is reduced in ER/PR positive and

negative breast tumors. Quantitative RT-PCR showing levels of HoxA9 mRNA in

normal human mammary tissue compared to levels in ER/PR negative and positive

human mammary tumor tissue. ***p<0.001 compared to normal mammary tissue

(normal: n=16, ER+/PR+ tumor: n=24, ER-/PR- tumor: n=23).

Supplemental Figure S2. Levels of HoxA9 protein upon re-expression in breast

tumor cell lines. Immunoblot showing level of expressed transgenic HoxA9 protein

attained in MDA-231 (left) and T4-2 (right) breast tumor cells compared to cells

expressing a vector control.

Supplemental Figure S3. Tetracycline regulated HoxA9 re-expression in breast

tumor cell lines. Epi-fluorescence microscopy images of breast tumor cells (MDA-231,

A & B; T4-2, A' & B') stably re-expressing retroviral HoxA9 bi-cistronically with EGFP,

showing transgene expression in the absence (B & B') and its loss (A & A') upon

tetracycline exposure (0.5 mg/ml; 72 hours). Insert: Phase contrast microscopy images of

A, A', B & B' indicating similar cell numbers were used under all experimental

conditions. Bars equal 50 mm.

Supplemental Figure S4. HoxA9 expression does not alter proliferation on 2D tissue

culture plastic. Growth curves of MDA-231 (black lines) and T4-2 (gray lines) breast

S13

tumor cells expressing a vector control (squares) or HoxA9 (circles) demonstrating that

exogenous HoxA9 expression does not alter 2D proliferation (n=3).

Supplemental Figure S5. HoxA9 re-expression phenotypically reverts breast tumor

cells. Bar graphs quantifying the tumor colony organization of T4-2 cells grown within a

3D reconstituted basement membrane for 10-12 days. Cell-cell junction integrity (left

graph) assessed by β-catenin cell-cell localization and continuity of staining. Basal

polarity (right graph) assessed by basal localization and continuity of β4 integrin

staining. ***P<0.001.

Supplemental Figure S6. Exogenous HoxA10 expression in MDA-231 breast tumor

cells. Semi-Q-PCR gel (left) and immunoblot (right) showing levels of HoxA10 RNA

and protein achieved upon re-expression in MDA-231 breast tumor cells. Lack of

chemiluminescent signal in HoxD10 lysates demonstrates the specificity of the HoxA10

antibody.

Supplemental Figure S7. HoxA10 expression in mammary epithelial tumor cells

does not reduce rBM colony growth. Phase contrast images of MDA-231 mammary

epithelial tumor cells expressing a vector control, HoxA9 or HoxA10 transgene showing

that growth inhibition within a 3D rBM is HoxA9 specific. Bar=50 μm.

Supplemental Figure S8. Mutational analysis of putative HoxA9 binding sites.

Luciferase reporter analysis showing continued responsiveness of BRCA1 promoter

S14

constructs to addition of wild-type HoxA9 when single putative HoxA9 binding sites are

deleted (Δ-221 to -218, Δ-175 to -172, and Δ-12 to -9) that is comparable to the

activation of the full length BRCA1 promoter construct (compare gray bars). Data are

normalized to matched vector control (black bars). Negative numbers refer to basepairs

upstream of the BRCA1 transcription start site (14).

Supplemental Figure S9. shRNA mediated HoxA9 knockdown. Q-RT-PCR analysis

of HoxA9 RNA levels in nonmalignant MCF10A mammary epithelial cells expressing a

luciferase control shRNA construct (Luc) or a HoxA9 shRNA clone.

Supplemental Figure S10. The BRCA1 Δexon11b mutant functions as a dominant

negative mutant. Immunoblot demonstrating that exogenous expression of BRCA1

Δexon11b (140KDa) in nonmalignant MCF10A cells completely abrogates levels of

wild-type BRCA1 (220KDa) suggesting the BRCA1 mutant functions as a dominant

negative.

S15

0

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

S2

[Ho

xA

9cD

NA

]

(ng

/L)]

m

[5’azacytidine]

0 Mm 1 Mm 10 Mm

Breast Normal 1 vs. Breast Cancer 1

Breast Normal 2 vs. Breast Cancer 2

Breast Normal 3 vs. Breast Cancer 3

Breast Normal 4 vs. Breast Cancer 4

Breast Normal 5 vs. Breast Cancer 5

0.2

0.4

0.6

0.8

-2 0 +2

0

1 0 0 0 0

2 0 0 0 0

3 0 0 0 0

S1

MCF10

AB

T-20

MD

A-4

68

MD

A-4

35

T47

D

ZR

751

MCF7

T4-2

MD

A-2

31

SK

BR

3

Re

lati

ve

In

ten

sit

y Tumor

S3

+Tet

-Tet

MDA 231 T4-2

A’

B’

A

B

S2

HoxA9

Actin

MDA-231

Vect

or

HoxA

9

0

25

50

75

100

Ba

sa

l P

ola

rity

Vecto

r

Rev

HoxA

9

T4-2 3D rBM

******

S5

0

25

50

75

100

Ce

ll-C

ell J

un

cti

on

Inte

gri

ty

Vecto

r

Rev

HoxA

9

T4-2 3D rBM

******

Supplemental Figure S1. HoxA9 expression is reduced in ER/PR positive and negative breast tumors.

Supplemental Figure S2.

Supplemental Figure S4. HoxA9 expression does not alter proliferation on 2D tissue culture plastic.

Supplemental Figure S5. HoxA9 re-expression phenotypically reverts breast tumor cells

<

Quantitative RT-PCR showing levels of HoxA9 mRNA in normal humanmammary tissue compared to levels in ER/PR negative and positive human mammary tumor tissue. ***p<0.001 compared to normal mammary tissue (normal: n=16, ER+/PR+ tumor:n=24, ER-/PR- tumor: n=23).

Growth curves of MDA-231 (black lines) and T4-2 (gray lines) breast tumorcells expressing a vector control (squares) or HoxA9 (circles) demonstrating that exogenous HoxA9 expression does not alter 2D proliferation (n=3).

. Bar graphs quantifying the tumor colony organization

***P 0.001.

Levels of HoxA9 protein upon re-expression in breast tumor cell lines.

Supplemental Figure S3. Tetracycline regulated HoxA9 re-expression in breast tumor cell lines.

Immunoblot showing level of expressed transgenic HoxA9 protein attained inMDA-231 (left) and T4-2 (right) breast tumor cells compared to cells expressing a vector control.

Epi-fluorescence microscopy images of breast tumor cells (MDA-231, A & B;T4-2,A' & B') stably re-expressing retroviral HoxA9 bi-cistronically with EGFP, showing transgene expression in the absence (B & B') and its loss (A&A') upon tetracycline exposure

(0.5 g/ml; 72 hours). Insert: Phase contrast microscopy images ofA,A', B & B' indicating similar cell numbers were used under all experimental conditions. Bars equal 50 m.

of T4-2 cells grown within a 3D

reconstituted basement membrane for 10-12 days. Cell-cell junction integrity (left graph) assessed by -catenin cell-cell localization and continuity of staining. Basal polarity (right

graph) assessed by basal localization and continuity of 4 integrin staining.

m m

b

b

T4-2

HoxA9

E-cadherin

Vect

or

HoxA

9

S1

0

1 0 0

2 0 0

Ho

xA

9 c

op

ies/

1e

7 c

op

ies 1

8S

Norm

al

ER

+/P

R+

ER

-/PR

-

******

Tumor

S4

0 1 2 3 4 5 6 7

0

300000

600000

900000

1200000

1500000

1800000

Time (days)

ave

rag

e #

of

ce

lls

S16

0 1 2 3

Vect

or

MT

BRCA1

MCF10A

220 KDa

140 KDa

S10

0

2 0

4 0

6 0

8 0

1 0 0

MCF10A

S9Lu

c

shR

NA

HoxA

9

shR

NA

#3

HoxA

9

shR

NA

#5

%D

ecre

ase

in

Ho

xA

9E

xp

ressio

n

S7MDA-231 3D rBM

Vector HoxA9 HoxA10

HoxA10

18S

Vect

or

Hox

A10

S6

HoxA10

Vect

or

Hox

D10

Hox

A10

S8

Luc-221 to -218

Luc-175 to -172

Luc-12 to -9

LucFull Length

Relative Promoter Activity

**

**

**

*

Supplemental Figure S6. Exogenous HoxA10 expression in MDA-231 breast tumor cells

Supplemental Figure S7. HoxA10 expression in mammary epithelial tumor cells does not reduce rBM colony growth=

Supplemental Figure S8. Mutational analysis of putative HoxA9 binding sites.

Supplemental Figure S9. shRNA mediated HoxA9 knockdown.

Supplemental Figure S10. The BRCA1 exon11b mutant functions as a dominant negative mutant.

. Semi-Q-PCR gel (left) and immunoblot (right) showing levels of HoxA10 RNA andprotein achieved upon re-expression in MDA-231 breast tumor cells. Lack of chemiluminescent signal in HoxD10 lysates demonstrates the specificity of the HoxA10 antibody.

. Phase contrast images of MDA-231 mammary epithelialtumor cells expressing a vector control, HoxA9 or HoxA10 transgene showing that growth inhibition within a 3D rBM is HoxA9 specifi

Luciferase reporter analysis showing continued responsiveness of BRCA1 promoter constructs toaddition of wild-type HoxA9 when single putative HoxA9 binding sites

rs upstream of the BRCA1 transcriptionstart site (9).

Q-RT-PCR analysis of HoxA9 RNA levels in nonmalignant MCF10A mammary epithelial cells expressing aluciferase control shRNAconstruct (Luc) or a HoxA9 shRNAclone.

Immunoblot demonstrating that exogenous expression of(140KDa) in nonmalignant MCF10Acells completely abrogates levels of wild-type BRCA1 (220KDa) suggesting the BRCA1 mutant functions as a dominant negative.

c. Bar 50 μm.

are deleted (Δ-221 to -218, Δ-175 to -172, and Δ-12 to -9) that is comparable to the activation of the full lengthBRCA1 promoter construct (compare gray bars). Data are normalized to matched vector control (black bars). Negative numbers refer to basepai

BRCA1 Δexon11bΔ

MDA-231 MDA-231

S17